How We Forecast Pollen at Planetary Scale

devdavidkarlsson — Sat, 09 May 2026 19:44:27 +0000

Most pollen apps show you a single number: "high." That's not very useful if you're allergic to birch but not oak. We built Atmospore to forecast pollen at the species level, globally,
updated daily. Here's how it works.

Why 25 km resolution is the right call

Our model runs at roughly 25 km between grid points. That might sound coarse compared to a street-level air quality sensor. But pollen doesn't work like car exhaust.

Airborne pollen concentrations are governed by mesoscale weather: wind fields, frontal passages, precipitation, and temperature gradients that operate over tens of kilometres. Our grid captures exactly these
drivers.

In a peer-reviewed study, researchers ran an atmospheric transport model at 36 km resolution across the entire US and found it successfully reproduced observed seasonal pollen distributions (Pearson r =
0.35–0.40 against 58 monitoring stations).[1]

Sub-kilometre variation does exist within cities. One study documented up to 300% differences across Berlin.[2] But that variation comes from individual trees and local wind channelling, not from anything a
forecast can predict. Everything finer would be noise dressed up as signal.

How pollen moves through the atmosphere

A birch pollen grain is roughly 22 micrometres in diameter and weighs around 5–10 nanograms. Once released, it behaves like a fine aerosol particle. Wind carries it horizontally while gravitational settling and
turbulent diffusion control its vertical profile.

On a dry, windy spring morning, birch pollen can reach 2 km altitude and travel 100+ km before depositing. Wet deposition (rain washing pollen from the air) is the primary removal mechanism and can cut
concentrations by 80% within hours.[3]

Why species matter

Most people are allergic to specific species, not to pollen in general. Someone who reacts strongly to birch may tolerate oak without symptoms.

On top of that, individual pollen grains vary in how much allergen they carry. Research has shown that birch pollen from different regions can contain very different levels of the Bet v 1 protein that triggers
the immune response.[5] A total pollen count doesn't tell you much if you don't know which species are in the air.

That's why Atmospore forecasts at the species level: 25 species, any land point on Earth.

Validated against real observations

We continuously calibrate our models against a wide array of real measurements from across the continents. Each species model is finely tuned with state-of-the-art technology and continuously optimised.

The model is never static. Every season brings new observations, and we continuously improve the models. We model species individually so that the forecast tracks each one's distinct behaviors. The goal is
always the same: if you're allergic and it's in the air, we want the forecast to say so.

Try the API

The Atmospore API delivers these forecasts as structured JSON. Species-resolved, georeferenced, updated daily. Free tier trial available, no credit card required.

  curl "https://pollenapi.com/v1/pollen?lat=59.91&lon=10.75&dt=2026-05-09&species=birch&forecast_days=3" \
    -H "x-api-key: YOUR_KEY"

  Get a free API key

References

Ren, X. et al. (2022). Modeling past and future spatiotemporal distributions of airborne allergenic pollen across the contiguous United States. Frontiers in Allergy, 3. PMC9640548
Werchan, B. et al. (2017). Spatial distribution of allergenic pollen through a large metropolitan area. Environmental Monitoring and Assessment, 189(169). PMID: 28316024
Sofiev, M. et al. (2013). A numerical model of birch pollen emission and dispersion in the atmosphere. International Journal of Biometeorology, 57, 45-58.
Anderegg, W.R.L. et al. (2021). Anthropogenic climate change is worsening North American pollen seasons. PNAS, 118(7).
Buters, J.T.M. et al. (1997). Release of Bet v 1 from birch pollen. Journal of Allergy and Clinical Immunology. JACI